Re: Recovery at replica stuck because recovery incorrectly trusts an old high-water mark

From: Jeff Davis <pgsql(at)j-davis(dot)com>
To: Konstantin Knizhnik <knizhnik(at)garret(dot)ru>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: Recovery at replica stuck because recovery incorrectly trusts an old high-water mark
Date: 2026-09-24 21:32:38
Message-ID: 867adbba98743e09428e4051f5b9013b4f4f10e1.camel@j-davis.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Sun, 2026-09-13 at 16:30 +0300, Konstantin Knizhnik wrote:
> A physical standby can get stuck forever if local pg_wal contains a
> corrupt record that walreceiver has already reported as flushed.
> What happens:

I can't help but ask: how did it get corrupted? Is it also corrupted on
the primary? What kinds of failures can cause this situation?

There seems to already be some kind of policy that we should try to re-
fetch bad records. Are you extending that policy or filling in an
implementation gap?

Regards,
Jeff Davis

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Masahiko Sawada 2026-09-24 22:08:40 Re: Fix "unexpected logical decoding status change" error; from concurrent logical decoding activation
Previous Message Robert Haas 2026-09-24 21:02:07 Re: pg_*_advice: tsv load failure, etc.